Hybrid quantum and classical computation: exploiting the best of both paradigms

点击翻译按钮获取中文摘要

Digital electronic computation has become ubiquitous on a very rapid timescale: more and faster computation is in greater demand than ever. Quantum computing promises more raw computing power than we can achieve classically: turning this promise into reality is the overarching goal of my research. I will address the key theoretical issue that will enable us to fully exploit quantum computation: how to combine quantum and classical computation to gain maximum computational power and efficiency.It is a crucial time to step up the development of quantum computing: Google recently bought their first "quantum computer". This device, from D-Wave (Burnaby, Canada), is solving real problems for commercial applications, even though we don't yet know whether it is actually exploiting quantum mechanics to achieve efficient computation beyond the reach of classical machines. Quantum computing is clearly coming of age: to ensure the UK has a place in the forefront of these developments we need our theorists and experimentalists to play their part in leading this computing revolution.My research is central to the key questions the D-Wave quantum computer challenges us with.How, exactly, do we persuade quantum systems to solve hard classical problems efficiently for us? We are part of the way there, we already know how to solve quantum problems: Feynman in 1982 first described how a quantum computer could efficiently simulate quantum systems, and experiments that can do this are well under way in labs around the world. Classical problems are tougher, there are relatively few algorithms promising a speed up. To use a quantum computer to solve a classical problem, such as factoring large numbers, or searching a random data set, or finding the best solution under a complex set of constraints, or modelling a large system (climate or proteins for example), we need a hybrid classical-quantum device that can start with the classical problem, convert it into a quantum representation, solve it, and then return the solution as classical data. Existing theoretical models of computation are simple, elegant, single paradigm models that perform well for analysis of complexity and computability - how hard it is to solve, and what are the minimum resources required - but methods of combining different models into hybrid composites that more closely match real computational devices are missing. Even the simplest experimental quantum processor is a hybrid device, typically combining classical controling hardware with two or more different quantum systems interacting through precisely specified sequences of operations. Hybrid quantum systems enable more practical experiments and more efficient quantum computer programs, both of which are essential to reduce the noise that would otherwise render quantum devices useless. But we don't yet know what is the best model of computation we should use to physically build useful computers. Silicon-based digital technology is serving us well, but the bienniel doubling of classical computing power is reaching quantum limitations in how small the elemental components can be made, and a diversity of less conventional devices are invading the marketplace for our daily productivity and entertainment. Niches are opening up for many special purpose types of computer, of which quantum is one important example. I will address these key gaps in our knowledge by developing a theoretical understanding of composite quantum-classical computational devices with real-world constraints applied, and by detailed theoretical and computational modelling of hybrid quantum-classical systems to characterise their properties, computational power and the conditions required for their efficient operation. This will enable me to provide the science and leadership that will place the UK in a prime position to produce and exploit the technology in the new era of quantum computation.